Ctrl+K
Ctrl+K

Spanish (Latin America) MFA dictionary v2.0.0a#

@techreport{mfa_spanish_latin_america_mfa_dictionary_2022,
	author={McAuliffe, Michael and Sonderegger, Morgan},
	title={Spanish (Latin America) MFA dictionary v2.0.0a},
	address={\url{https://mfa-models.readthedocs.io/pronunciation dictionary/Spanish/Spanish (Latin America) MFA dictionary v2_0_0a.html}},
	year={2022},
	month={May},
}

G2P models
../../_images/full_logo_yellow.svg

Installation#

Install from the MFA command line:

mfa model download dictionary spanish_latin_america_mfa

Or download from the release page.

The dictionary available from the release page and command line installation has pronunciation and silence probabilities estimated as part acoustic model training (see Silence probability format and training pronunciation probabilities for more information. If you would like to use the version of this dictionary without probabilities, please see the plain dictionary.

Intended use#

This dictionary is intended for forced alignment of Spanish transcripts.

This dictionary uses the MFA phone set for Spanish, and was used in training the Spanish MFA acoustic model. Pronunciations can be added on top of the dictionary, as long as no additional phones are introduced.

Performance Factors#

When trying to get better alignment accuracy, adding pronunciations is generally helpful, especially for different styles and dialects. The most impactful improvements will generally be seen when adding reduced variants that involve deleting segments/syllables common in spontaneous speech. Alignment must include all phones specified in the pronunciation of a word, and each phone has a minimum duration (by default 10ms). If a speaker pronounces a multisyllabic word with just a single syllable, it can be hard for MFA to fit all the segments in, so it will lead to alignment errors on adjacent words as well.

Ethical considerations#

Deploying any Speech-to-Text model into any production setting has ethical implications. You should consider these implications before use.

Demographic Bias#

You should assume every machine learning model has demographic bias unless proven otherwise. For pronunciation dictionaries, it is often the case that transcription accuracy and lexicon coverage for the prestige variety modeled in this dictionary compared to other variants. If you are using this dictionary in production, you should acknowledge this as a potential issue.

IPA Charts#

Consonants#

Obstruent symbols to the left of are unvoiced and those to the right are voiced.

Manner

Labial

Labiodental

Dental

Alveolar

Alveopalatal

Palatal

Velar

Nasal
Occurrences:
20,399
Examples:
* lomas:
[l o m a s]
* morro:
[m o r o]
* slim:
[s l i m]
* oímos:
[o i m o s]
Occurrences:
33,217
Examples:
* cuán:
[k w a n]
* milan:
[m i l a n]
* manes:
[m a n e s]
* notro:
[n o t ̪ ɾ o]
Occurrences:
2,273
Examples:
* ñirre:
[ɲ i r e]
* chuño:
[t ʃ u ɲ o]
* dueño:
[d ̪ w e ɲ o]
* nievo:
[ɲ j e β o]
Occurrences:
2,452
Examples:
* incas:
[i ŋ k a s]
* vence:
[b e ŋ s e]
* cinco:
[s i ŋ k o]
* anj:
[a ŋ x]

Stop
Occurrences:
15,153
Examples:
* proa:
[p ɾ o a]
* pelvi:
[p e l β i]
* apura:
[a p u ɾ a]
* puaj:
[p w a x]
Occurrences:
6,139
Examples:
* berry:
[b e r i]
* bulín:
[b u l i n]
* brión:
[b ɾ j o n]
* bei:
[b e i]
Occurrences:
32,610
Examples:
* estro:
[e s t ̪ ɾ o]
* goteo:
[ɡ o t ̪ e o]
* tiré:
[t ̪ i ɾ e]
* taxi:
[t ̪ a ɣ s i]
Occurrences:
9,467
Examples:
* dormí:
[d ̪ o r m i]
* debes:
[d ̪ e β e s]
* dark:
[d ̪ a r ɣ]
* duma:
[d ̪ u m a]
Occurrences:
2,130
Examples:
* like:
[l a i c e]
* yuki:
[ɟ ʝ u c i]
* kiwi:
[c i w i]
* roque:
[r o c e]
Occurrences:
175
Examples:
* yipi:
[ɟ ʝ i p i]
* guía:
[ɟ i a]
* guija:
[ɟ i x a]
* yuyo:
[ɟ ʝ u ʝ o]
Occurrences:
21,411
Examples:
* poca:
[p o k a]
* luc:
[l u k]
* educo:
[e ð u k o]
* cayo:
[k a ʝ o]
Occurrences:
2,243
Examples:
* glose:
[ɡ l o s e]
* bang:
[b a ŋ ɡ]
* goya:
[ɡ o ʝ a]
* gould:
[ɡ u l d ̪]

Affricate
Occurrences:
2,511
Examples:
* chapó:
[t ʃ a p o]
* mecha:
[m e t ʃ a]
* fichó:
[f i t ʃ o]
* tachi:
[t ̪ a t ʃ i]
Occurrences:
524
Examples:
* yaqui:
[ɟ ʝ a c i]
* yate:
[ɟ ʝ a t ̪ e]
* yapé:
[ɟ ʝ a p e]
* yinn:
[ɟ ʝ i n]

Sibilant
Occurrences:
55,093
Examples:
* glosa:
[ɡ l o s a]
* zhang:
[s x a ŋ ɡ]
* lazio:
[l a s i o]
* ajax:
[a x a k s]
Occurrences:
284
Examples:
* tocha:
[t ̪ o t ʃ a]
* chang:
[t ʃ a ŋ ɡ]
* chato:
[t ʃ a t ̪ o]
* gwh:
[ɡ e u β e ð o β l e a t ʃ e]

Fricative
Occurrences:
6,945
Examples:
* field:
[f j e l d ̪]
* firmó:
[f i r m o]
* áfido:
[a f i ð o]
* fiel:
[f j e l]
Occurrences:
27,680
Examples:
* orado:
[o ɾ a ð o]
* wide:
[w i ð e]
* quedo:
[c e ð o]
* vídeo:
[b i ð e o]
Occurrences:
2,948
Examples:
* jeta:
[ç e t ̪ a]
* jica:
[ç i k a]
* giba:
[ç i β a]
* jesus:
[ç e s u s]
Occurrences:
2,003
Examples:
* zülle:
[s w ʝ e]
* yuyu:
[ɟ ʝ u ʝ u]
* yudo:
[ɟ ʝ u ð o]
* hayan:
[a ʝ a n]

Approximant
Occurrences:
3,291
Examples:
* tenue:
[t ̪ e n w e]
* www:
[b w u]
* where:
[w e ɾ e]
* hueco:
[w e k o]
Occurrences:
9,945
Examples:
* abrió:
[a β ɾ j o]
* alió:
[a l j o]
* bien:
[b j e n]
* oriol:
[o ɾ j o l]

Tap
Occurrences:
33,710
Examples:
* libró:
[l i β ɾ o]
* atar:
[a t ̪ a ɾ]
* prado:
[p ɾ a ð o]
* abren:
[a β ɾ e n]

Trill
Occurrences:
15,932
Examples:
* horta:
[o r t ̪ a]
* barra:
[b a r a]
* risas:
[r i s a s]
* roano:
[r o a n o]

Lateral
Occurrences:
21,829
Examples:
* menlo:
[m e n l o]
* shell:
[ʃ e l]
* álex:
[a l e k s]
* locas:
[l o k a s]

Vowels#

Vowel symbols to the left of are unrounded and those to the right are rounded.

Front

Near-Front

Central

Near-Back

Back

Close
Occurrences:
40,637
Examples:
* ayres:
[a i ɾ e s]
* tías:
[t ̪ i a s]
* multi:
[m u l t ̪ i]
* quise:
[c i s e]
Occurrences:
14,204
Examples:
* juntó:
[x u n t ̪ o]
* púa:
[p u a]
* tubey:
[t ̪ u β e i]
* muñón:
[m u ɲ o n]

Close-Mid
Occurrences:
59,310
Examples:
* maite:
[m a i t ̪ e]
* cedro:
[s e ð ɾ o]
* señar:
[s e ɲ a ɾ]
* bagel:
[b a ʝ e l]
Occurrences:
58,466
Examples:
* cardo:
[k a r ð o]
* acoso:
[a k o s o]
* yokut:
[ɟ ʝ o k u t ̪]
* avión:
[a β j o n]

Open-Mid

Open
Occurrences:
98,488
Examples:
* solas:
[s o l a s]
* misa:
[m i s a]
* laves:
[l a β e s]
* salid:
[s a l i ð]
On this page
Show Source